JP5151786B2 - Storage element and storage method - Google Patents

Description

  The present invention relates to a storage element suitable for application as an image memory.

  In particular, the present invention relates to a semiconductor memory element capable of realizing a two-port operation of writing and reading while suppressing an increase in chip area.

  FIG. 1 shows an example of the configuration of a conventional image memory 100.

  The image memory 100 includes a DRAM (dynamic RAM) 101 and SRAMs (static RAMs) 102A and 102B that are arranged on the data input side (serial input side) and used as buffer memories.

  The memory 100 is also arranged as a serial Y decoder 103 that selects a write address for the SRAMs 102A and 102B based on the Y-direction component of the write address signal, and is arranged on the data output side (serial output side) and used as a buffer memory. SRAMs (static RAMs) 104A and 104B, and a serial Y decoder 105 that selects a read address for the SRAMs 104A and 104B based on the Y-direction component of the read address signal.

  Now, the SRAMs 102A and 102B on the data input side are arranged as half of the array of the DRAM 101, and perform transfer operations alternately so that data can be written sequentially.

  Similarly, the SRAMs 104A and 104B on the data output side are also arranged as half of the array of the DRAM 101, and alternately execute transfer operations so that data can be read sequentially.

  Further, the image memory 100 generates an X decoder 106 for selecting an X direction address (transfer address) related to writing or reading with respect to the DRAM 101, a write address counter 107 for generating a write address signal, and a read address signal. A read address counter 108, and an arbiter 109 used as an arbitration circuit for delaying the write transfer when the write transfer command and the read transfer command are close to each other.

  In this case, the write transfer itself may be performed until the next data write to the SRAM (buffer memory) is completed, and there is a sufficient lead time for the write transfer.

  For example, the image memory 100 shown in FIG. 1 is used for performing flicker-free signal processing for suppressing screen flicker (flicker) with, for example, twice the number of screens.

  FIG. 2 shows address changes between writing and reading in the non-flicker signal processing. In this case, each video data (1W, 2W, etc.) constituting one field or frame is written in succession in the image memory 100.

  Furthermore, each video data (1R, 2R, etc.) is continuously read out twice from the image memory 100 at twice the writing speed.

  Since the image memory 100 shown in FIG. 1 includes the buffer memories 102A, 102B, 104A, and 104B therein, there is no problem even when the write transfer command and the read transfer command are connected to each other. A two-port operation for writing and reading can be performed.

  However, since the sizes of the buffer memories 102A, 102B, 104A and 104B are large, the chip area of the image memory 100 increases, which is disadvantageous.

  In recent years, due to restrictions on the extension of a word and bit line, a megabit class memory is often divided into a plurality of memory blocks (MAT unit, modular arrangement unit) each serving as an individual memory. It is.

  In this case, when the read and write operations are performed on the scheduled memory block, the memory block that is not accessed is put in an inactive state.

  The present invention has been made in view of the above situation, and one object of the present invention is to provide a memory element capable of performing two-port operations for writing and reading without increasing the chip area. Is to provide.

  Another object of the present invention is to provide a storage element capable of performing a synchronized two-port operation for writing and reading without using a buffer memory.

In order to solve such a problem, in the present invention, a storage unit including N (N is an even number) memory blocks that can function individually, and M blocks (N>M> N / 2, M are A write address signal for sequentially writing each image data having a size equal to ( even number) into the storage unit, and a predetermined number of bits counted from the most significant bit in which block in the image data a write address generator for generating a write address signal indicating whether the corresponding to the image data, Ru is output from the write address generator when the writing of the image data is finished to have a size equal to M blocks digits Whether the memory block to be written for each image data is shifted by M blocks based on the up signal A write control unit for writing on the basis of the write address signal, and a read address signal for sequentially reading out each image data from the storage unit , and a predetermined number of bits counted from the most significant bit. A read address generator that generates a read address signal indicating the number of blocks in the image data at twice the speed of the write address signal generated by the write address generator, and image data written in the storage unit , the memory block to be read on the basis of once every two carry signals from the read address generator Ru is output when the reading of image data having a size equal to M blocks is finished for each image data Based on read address signal from memory block shifted by M blocks And a reading control unit to read.

As a result, a two-port operation of writing and reading can be executed, while an increase in the area of the chip is suppressed.

In addition, in a storage unit including N (N is an even number) memory block that can function individually, a size equal to M blocks (N>M> N / 2, where M is an even number) in order. a write address signal for write the respective image data having a writing a write address signal indicating whether a predetermined number of bits counted from the most significant bit corresponds to the ordinal number of the block in the image data and generating by the address generator, the image data for each image data based on the carry signal from the write address generator Ru is output when the writing of the image data is finished to have a size equal to M blocks Write from the memory block in which the memory block to be written is shifted by M blocks based on the write address signal And a read address signal for sequentially reading out each image data from the storage unit, which number block in the image data corresponds to a predetermined number of bits counted from the most significant bit A read address signal to be generated by the read address generator at a rate twice as fast as the write address signal, and the image data written in the storage unit is read out of image data having a size equal to M blocks a memory block of M blocks only the shifted memory block to be read for each image data based on once every two carry signals to be output from the read address generator when but ended, based on the read address signal And a step for reading.

As a result, a two-port operation of writing and reading can be executed, while an increase in the area of the chip is suppressed.

According to the present invention, the storage unit includes N memory blocks, each functioning as a separate memory. When the data in the memory block accompanied by M (M <N) data as a unit is sequentially written into the storage unit and continuously read out from it, the writing and reading of each unit described above is started. Since the address is shifted in the memory block as a unit, writing and reading are not performed simultaneously in the same memory block.

Further, in the present invention, a predetermined number of revised bits for selecting a memory block from which image data is to be read is replaced, and the image data is written in a predetermined number of bits counted from the most significant bit of the write address signal. Replace with the planned number of revised bits to select the memory block to be.

As a result, a two-port operation of writing and reading can be executed, while an increase in the area of the chip is suppressed.

  A first embodiment of the present invention will be described below with reference to the accompanying drawings.

  FIG. 3 shows the configuration of the semiconductor memory element 10 used as the first embodiment.

  The semiconductor memory element 10 of the first embodiment is for executing a synchronous two-port operation without the need to use a buffer memory, and is a flicker-free signal that prevents screen flickering with twice the number of screens. Intended to achieve processing.

  The storage element 10 includes a storage element unit 11 composed of eight memory blocks (DRAM arrays) 12a to 12h that function as separate memories.

  Each of the memory blocks 12a to 12h is configured to have a 128 kilogram DRAM, an X decoder for selecting an address in the X direction of the DRAM, and a Y decoder for selecting an address in the Y direction of the DRAM. Is done.

  In this case, the video data per field or frame will be equal to 6 data in each memory block.

  Further, the storage element 10 includes four selectors 13a to 13d for selecting the memory blocks 12a to 12h as objects to be written or read.

  In this case, the selectors 13a to 13d are configured to select two memory blocks as a pair.

  The selector 13a selects the memory blocks 12a and 12b as a pair, the selector 13b selects the memory blocks 12c and 12d as a pair, the selector 13c selects the memory blocks 12e and 12f as a pair, and the selector 13d selects the memory blocks 12g and 12h. Are selected as a pair.

  In this way, two memory blocks in each pair are selected by each of the selectors 13a to 13d in order to enable continuous serial operation by bank switching.

  The storage element 10 includes a write address counter 14, a write address pointer 15, and a difference element 16.

  The write address counter 14 generates a 20-bit write address signal.

  The write address pointer 15 counts up in accordance with the carry signal output from the write address counter 14 and outputs a 2-bit signal therefrom.

  The difference element 16 is a 2-bit signal output from the pointer 15 from an upper 2-bit signal (MSB (most significant bit) and second bit from the MSB) of the write address signal generated from the write address counter 14. Reduce.

  In this case, the high-order 2-bit signal of the 20-bit write address signal generated from the counter 14 is replaced with the 2-bit signal output from the difference element 16 to create or form the write address signal WAD for the storage unit 11. Is done.

  In this case, a pair of memory blocks to be written is selected based on the upper 2 bits (MSB and the second bit from the MSB) of the write address signal WAD.

  The storage element 10 includes a read address counter 17, a read address pointer 18, and a difference element 19.

  The read address counter 17 generates a 20-bit read address signal.

  The read address pointer 18 counts up according to the carry signal output from the read address counter 17 and outputs a 2-bit signal therefrom.

  The difference element 19 outputs the 2-bit signal output from the pointer 18 from the upper 2-bit signal (MSB (most significant bit) and the second bit from the MSB) of the read address signal generated from the read counter 17. Decrease.

  In this case, the upper 2-bit signal of the read address signal generated from the counter 17 is replaced with the 2-bit signal output from the difference element 19 to form the read address signal RAD for the storage unit 11.

  In this case, a pair of memory blocks to be read is selected based on the upper 2 bits (MSB and the second bit from the MSB) of the read address signal RAD.

  The count speed, that is, the reading speed of the counter 17 is set to twice the count speed of the counter 14, that is, the reading speed.

  An upper 2-bit signal of a read address signal RAD described later, that is, a 2-bit signal obtained by adding 1 to the 2-bit signal output from the difference element 19, is an upper 2-bit signal of the write address write address signal WAD, The pointer 18 is counted up according to the carry signal output from the counter 17 only when it is not equal to the 2-bit signal output from the difference element 16.

  As described above, since the read start address is changed every time reading is performed twice, video data for the same one field or one frame is continuously read.

Next, the operation of the memory element 10 will be described with reference to FIGS . 4 (A) and 4 (B) . FIG. 4A shows addresses on the image.

Each solid line in FIG. 4 (A) shows the change of the write address of the image, the broken lines in FIG. 4 (A) shows the change of the read address on the image.

  That is, the write address corresponds to the write address signal generated from the counter 14, and the read address corresponds to the read address signal generated from the counter 17.

4A and 4B , 1W, 2W, and 3W respectively indicate video data for one field or one frame related to writing.

  Reference numerals 1R, 2R, and 3R indicate video data for one field or one frame related to reading, and correspond to the video data 1W, 2W, and 3W, respectively.

  If the write address and read address on the image are used as they are, their upper 2 bit signals change in the order of “00” → “01” → “10” → “00” → “01” →. To do.

  Only the pair of memory blocks 12a and 12b, the pair of memory blocks 12c and 12d, and the pair of memory blocks 12e and 12f are used for writing and reading. In the range of Q, since writing and reading are simultaneously performed in the same pair of memory blocks, it is not possible to realize a two-port operation for writing and reading.

In the present invention, the writing and reading of the addresses on the memory are employed as shown in FIG. 4 (B).

FIG. 4B shows write and read addresses on the memory.

Each solid line in FIG. 4 (B) shows a change in write address on the memory, the broken line in FIG. 4 (B) shows a change in read address on the memory.

  The write address corresponds to the write address signal WAD, and the read address corresponds to the read address signal RAD in FIG.

  In this case, the upper 2 bits signal of the write address signal WAD changes in the order of “00” → “01” → “10” with respect to the video data 1W.

  Video data 1W is sequentially written into a pair of memory blocks 12a and 12b, a pair of memory blocks 12c and 12d, and a pair of memory blocks 12e and 12f.

  Further, the upper 2 bits signal of the read address signal RAD changes in the order of “00” → “01” → “10” → “00” → “01” → “10” with respect to the video data 1R.

  Then, the video data 1R is read twice from the pair of the memory blocks 12a and 12b, the pair of the memory blocks 12c and 12d, and the pair of the memory blocks 12e and 12f sequentially at twice the writing speed. .

  Next, the upper 2 bits signal of the write address signal WAD changes in the order of “11” → “00” → “01” with respect to the video data 2W following the video data 1W.

  Video data 2W is sequentially written into a pair of memory blocks 12g and 12h, a pair of memory blocks 12a and 12b, and a pair of memory blocks 12c and 12d.

  Further, the upper 2 bits signal of the read address signal RAD changes in the order of “11” → “00” → “01” → “11” → “00” → “01” with respect to the video data 2R.

  The video data 2R is continuously transmitted twice from the pair of the memory blocks 12g and 12h, the pair of the memory blocks 12a and 12b, and the pair of the memory blocks 12c and 12d at a speed twice as high as that at the time of writing. Is read out.

  Thereafter, when the write operation and the read operation are performed in the same manner as described above, and when regular video data is supplied from the data input side (serial input side), the number of screens is reduced to prevent screen flicker. Video data for setting to double is obtained on the data output side (serial output side).

  In the first embodiment as described above, the start address for writing and reading is sequentially shifted, thereby preventing the writing and reading from being performed simultaneously in the same pair of memory blocks, and for the writing and reading. Realizes 2-port operation.

  Thus, since this embodiment does not use a large buffer memory, an increase in chip area can be suppressed.

FIG. 5 shows a semiconductor memory element 200 used as the second embodiment of the present invention.

  The storage element 200 realizes a synchronous two-port operation without using a buffer memory.

  More specifically, the storage element 200 performs noise reduction signal processing by accessing a memory block free from writing and reading operations.

  The semiconductor memory element 200 as the second embodiment is for realizing a synchronous two-port operation without being forced to use a buffer memory, and reduces noise by using an inactive area that is not accessed. It is intended to perform signal processing.

  The storage element 200 has a storage unit 211 composed of ten memory blocks (DRAM arrays) 212a to 12j each functioning as an individual memory.

  Each of these memory blocks 212 to 212j includes a 128 kilogram DRAM, an X decoder for selecting an address in the X direction of the DRAM, and a Y decoder for selecting an address in the Y direction of the DRAM. Configured.

  In this case, the video data per field or frame will be equal to 6 data in each memory block.

  Further, the storage element 200 includes five selectors 213a to 213e for selecting the memory blocks 212a to 212j as objects to be written or read.

  In this case, the selectors 213a to 213e are configured to select two memory blocks as a pair.

  The selector 213a selects the memory blocks 212a and 212b as a pair, and so on.

  As described above, each of the selectors 213a to 213e selects two memory blocks in each pair, thereby enabling continuous serial operation by bank switching.

  The storage element 200 includes a write address counter 214, a write address pointer 215, and a lookup table 216.

  The write address counter 214 generates a 20-bit write address signal.

  The write address pointer 215 counts up according to the carry signal output from the write address counter 214 and outputs a 3-bit signal therefrom.

  The look-up table 216 includes a 3-bit signal output from the pointer 215, an upper 2-bit signal (MSB (most significant bit) of the write address signal generated from the write address counter 214, and a second bit from the MSB. ) To a 3-bit output.

  In this case, the upper 2-bit signal of the 20-bit write address signal generated from the counter 214 is replaced with the 3-bit signal output from the lookup table 216, so that the write address signal WAD for the storage unit 211 is created or It is formed.

  In this case, a pair of memory blocks to be written is selected based on the upper 3 bits (MSB, the second bit from the MSB, and the third bit from the MSB) of the write address signal WAD.

  The storage element 200 includes a read address counter 217, a read address pointer 218, and a lookup table 219.

  The read address counter 217 generates a 20-bit read address signal.

  The read address pointer 218 counts up according to the carry signal output from the read address counter 217 and outputs a 3-bit signal therefrom.

  The look-up table 219 includes a 3-bit signal output from the pointer 218, an upper 2-bit signal (MSB (most significant bit), and the second bit from the MSB) of the read address signal generated from the read address counter 217. Is converted to a 3-bit output.

  In this case, the upper 2-bit signal of the read address signal generated from the counter 217 is replaced with the 3-bit signal output from the look-up table 219, and the read address signal RAD for the storage unit 211 is formed.

  In this case, a pair of memory blocks to be read is selected based on the above three bits (MSB, the second bit from the MSB, and the third bit from the MSB) of the read address signal RAD.

  The storage element 200 further includes an IIR read address counter 317, an IIR read address pointer 318, and an IIR lookup table 319.

  The IIR read address counter 317 generates a 20-bit read address signal.

  The IIR read address pointer 318 counts up according to the carry signal output from the IIR read address counter 317 and outputs a 3-bit signal therefrom.

  The IIR lookup table 319 includes a 3-bit signal output from the pointer 318 and an upper 2-bit signal (MSB (most significant bit) and the second bit from the MSB) of the read address signal generated from the read address counter 317. And convert.

  These operate in the same manner as previously described to allow the IIR (Infinite Impulse Response) filter circuit 300 to access the storage unit 211.

  The IIR filter circuit 300 sequentially obtains access to the memory block 211 to obtain input video data and video data created by delaying the input video data by one field or frame, thereby Perform noise reduction processing.

  Furthermore, this IIR filter circuit writes video data from which noise has been removed into each memory block instead of the above-described input video data.

FIG. 6 shows the look-up tables 216, 219 and 319. The left three columns show the input from the 3-bit input coming from pointers 215, 218 and 318. The middle two columns show the 2-bit input from counters 214, 217 and 317. The right three columns show the 3-bit output from lookup tables 216, 219 and 319.

  FIG. 6 shows a configuration of a known IIR type noise reduction circuit 300 as in the past.

  Noise reduction circuit 300 includes a coefficient multiplier 320, a coefficient multiplier 330, an adder 340, a coefficient multiplier 350, a subtractor 360, and a coefficient output circuit 370.

  The coefficient multiplier 320 multiplies the input video data Vin by the coefficient k.

  The coefficient multiplier 330 multiplies the video data Vdo output from the memory block 211 by one field or one frame before by the coefficient 1-k.

  The adder 340 and the coefficient multiplier 350 add and average the output video data of the coefficient multiplier 320 and the coefficient multiplier 330 to obtain output video data Vout from which noise is removed. The subtracter 360 and the coefficient output circuit 370 perform subtraction between the input video data Vin and the video data Vdo output from the memory block 211, and a coefficient k (0 <k) corresponding to a motion by a signal representing the subtraction. <1) for obtaining.

  The output video data is output as video data from which noise has been removed, and is stored in the storage unit 211.

Referring to FIG. 5 again, the IIR filter circuit 300 performs noise reduction processing to obtain output video data Vout from which noise has been removed.

  In this case, the IIR filter circuit 300 does not need to have a memory.

  That is, the IIR filter circuit 300 obtains the input video data Vin and the output video data Vdo positioned one field before by acquiring access to the plurality of memory blocks constituting the storage unit 211. Incidentally, the noise reduction processing by the IIR filter circuit 300 is performed before reading out video data from which noise has not been removed.

Next, the operation of the memory element 200 will be described with reference to FIG. 7 (A) and 7 (B).

FIG. 7A shows addresses on the image.

Each solid line in FIG. 7 (A) shows a change in write address on the image, the broken lines in FIG. 7 (A) shows a change in read address on the image.

  This write address corresponds to a write address signal generated from the counter 214, and a read address indicated by a broken line corresponds to a read address signal generated from the counter 217.

The counter 317 generates a read address signal IIR as each hatched as shown in FIG. 7 (B).

7A and 7B , 1 W, 2 W, 3 W,... Respectively indicate video data for one field or one frame related to writing.

  1R, 2R and 3R,... Indicate video data for one field or one frame related to reading, and correspond to the video data 1W (1C), 2W (2C) and 3W (3C). doing. ICs 2C and 3C also show video data for one field or one frame related to reading, and correspond to the video data 1W, 2W and 3W, respectively.

  In this case, the upper 3 bits of the write address signal WAD change in the order of “000” → “001” → “010” →.

  Then, the video data 1W is sequentially written into the pair of memory blocks 212a and 212b, the pair of memory blocks 212c and 212d, and the like without noise reduction processing.

  This is because there is no previous video data in the storage unit 211.

  Further, the upper 3 bits signal of the read address signal RAD changes in the order of “000” → “001” → “010” with respect to the video data 1C.

  The video data 1C is sequentially read from the pair of memory blocks 212a and 212b, the pair of memory blocks 212c and 212d, and the pair of memory blocks 212e and 212f to be supplied to the IIR filter circuit 300.

  Thereafter, in the IIR filter circuit 300, the input video data 2W following the video data 1W is subjected to noise reduction processing using the video data 1C from the storage unit 211, and the video data 2W from which noise has been removed is converted into IIR. Output from the filter circuit 300.

  In this case, the video data written in the storage unit 211 is replaced with video data (output video data Vout) from which noise has been removed by the IIR filter circuit 300.

  At the same time, the upper 3 bid signal of the write address signal WAD changes in the order of “011” → “100” → “000” with respect to the video data 2W from which noise has been removed following the video data 1W.

  The video data 2W from which noise has been removed is sequentially written into a pair of memory blocks 212g and 212h, a pair of memory blocks 212i and 212j, and a pair of memory blocks 212a and 212b.

  Further, the upper 3 bits of the read address signal RAD change in the order of “000” → “001” → “010” with respect to the video data 1R, and the video data 1R includes the memory blocks 212a and 212b and the memory blocks 212c and 212c. The data is continuously read from the pair 212d and the pair of memory blocks 212e and 212f and output from the symbol element 200.

  Next, the upper 3 bits signal of the read address signal IIR changes in the order of “011” → “100” → “000” with respect to the video data 2C from which noise has been removed.

  Then, the video data 2C from which noise has been removed is sequentially read from the pair of memory blocks 212g and 212h, the pair of memory blocks 212i and 212j, and 212a and 212b in order to be supplied to the IIR filter circuit 300.

  Then, in the IIR filter circuit 300, noise reduction processing is performed on the input video data 3W following the video data using the video data 2C from the storage unit 211, and the video data 3W from which noise has been removed is converted into an IIR filter. Output from the circuit 300.

  In this case, the video data written in the storage unit 211 is replaced with video data (output video data Vout) from which noise has been removed by the IIR filter circuit 300.

  At the same time, the upper 3 bits of the write address signal WAD change in the order of “011” → “100” → “000” with respect to the video data 2R from which noise has been removed.

  The noise-removed video data 2R is continuously read from the pair of memory blocks 212g and 212h, the pair of memory blocks 212i and 212j, and the pair of memory blocks 212a and 212b. Is output.

  When the write operation and the read operation are performed in the same manner as described above and normal video data is supplied from the data input side (serial input side), noise is generated on the data output side (serial output side). Is obtained.

  Video data obtained on the data output side is dominated by noise reduction processing.

  In the second embodiment described above, the two-port operation for writing and reading is performed by sequentially shifting the starting address of writing and reading so that writing and reading are not performed simultaneously in the same pair of memory blocks. Is realized.

  Thus, this embodiment does not use a large buffer memory, and can suppress an increase in the chip area in the same manner as the first embodiment.

  Furthermore, the IIR filter 300 performs noise reduction signal processing by acquiring access to a memory block that is not involved in write and read operations.

  This makes it possible to effectively use an inactive memory area that has not been accessed.

  Although the IIR filter circuit 300 for performing noise reduction processing is provided in the second embodiment described above, an IIR that acquires access to a memory block that is not involved in write and read operations is provided and preset. The obtained video data may be obtained, thereby performing other processing.

  For example, a resolution creation process (see Japanese Patent Application No. 6-205934) for converting NTSC video data into high-definition television image data, and a progressive conversion process for converting interlaced video data into non-interlaced video data The motion vector detection process for MPEG (moving picture compression encoding system), etc. can be applied as other processing.

  According to the present invention, the storage unit includes N memory blocks, each functioning as a separate memory.

  When the data in the memory block accompanied by M (M <N) data as a unit is sequentially written into the storage unit and continuously read out from it, the writing and reading of each unit described above is started. Since the address is shifted in the memory block as a unit, writing and reading are not performed simultaneously in the same memory block.

  As a result, a two-port operation of writing and reading can be executed, while an increase in the area of the chip is suppressed.

  Furthermore, according to the present invention, a two-port operation for writing and reading can be performed without a buffer memory.

  Thus, it can be realized with a small-sized chip.

  Moreover, according to the present invention, a plurality of memory blocks each functioning as an independent memory, and an arithmetic circuit for acquiring access to a memory block free from writing and reading operations and thereby performing signal processing And a storage unit is provided.

  In this way, there is a gain in that an unaccessed bidirectional area can be effectively used.

  Although the invention has been fully described herein, it will be apparent to those skilled in the art that many changes and modifications can be made thereto without departing from the spirit and scope of the invention described herein.

To assist in a more complete understanding of the invention, the following description and accompanying drawings are set forth.
FIG. 1 is a block diagram showing an example of the configuration of a conventional image memory. FIG. 2 is a diagram showing address changes during flickerless signal processing achieved using a conventional image memory. FIG. 3 is a block diagram showing the configuration of the semiconductor memory element used as the first embodiment. FIGS. 4 (A) and. 4 (B), for explaining the first embodiment, and shows the address change when no flicker signal processing. FIG. 5 is a block diagram showing a configuration of a semiconductor memory element used as the second embodiment. FIG. 6 is a table showing a lookup table (LUT). FIGS. 7A and 7B are diagrams showing address changes during IIR filter processing for explaining the second embodiment. FIG. 8 is a block diagram of the IIR filter.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Semiconductor memory element, 11 ... Memory unit, 12a-12h ... Memory block, 13a-13d ... Selector, 14 ... Write address counter, 15 ... Write address pointer, 16, 19 ... Difference element, 17: Read address counter, 18: Read address pointer.

Claims (6)

A storage unit including N (N is an even number) memory blocks each capable of functioning individually;
A write address signal for sequentially writing each image data having a size equal to M blocks (N>M> N / 2, M is an even number) into the storage unit. A write address generator for generating a write address signal indicating the number of blocks in the image data corresponding to a predetermined number of bits counted from the bits;
The image data, the memory block to be written for each image data based on the carry signal that will be output from the write address generator when the writing of the image data having a size equal to M blocks is completed A write control unit for writing based on the write address signal from a memory block shifted by M blocks;
A read address signal for sequentially reading the image data from the storage unit , which indicates which block in the image data corresponds to a predetermined number of bits counted from the most significant bit A read address generator that generates an address signal at twice the speed of the write address signal generated by the write address generator;
The image data written in the memory unit, once the two carry signals that will be output from the read address generator when the reading of image data having a size equal to M blocks is completed And a read control unit for reading from a memory block obtained by shifting a memory block to be read for each image data by M blocks based on the read address signal. Each image data is
The storage element according to claim 1, wherein the storage element represents image data for one field or one frame. The storage element according to claim 2, further comprising: an arithmetic circuit that performs certain signal processing by acquiring access to a memory block in the storage unit that is not involved in the write and read operations. The storage element according to claim 3, wherein the arithmetic circuit performs noise reduction processing. The arithmetic circuit is
Receives the current image data from the input terminal, receives previous image data from the memory block not involved in the writing and reading operations, and uses the previous image data to The storage element according to claim 4, wherein the noise reduction processing is performed to generate current image data with reduced noise. A storage unit including N memory blocks (N is an even number) that can function individually has a size equal to M blocks (N>M> N / 2, where M is an even number) in sequence. a write address signal for write the respective image data, and write address generator to the write address signal indicating whether a predetermined number of bits counted from the most significant bit corresponds to the ordinal number of the block in the image data The step of generating by
The image data, the memory block to be written for each image data based on the carry signal that will be output from the write address generator when the writing of the image data having a size equal to M blocks is finished M Writing from the memory block shifted by the number of blocks based on the write address signal;
A read address signal for sequentially reading the image data from the storage unit , which indicates which block in the image data corresponds to a predetermined number of bits counted from the most significant bit Generating an address signal by a read address generator at a rate twice that of the write address signal;
The image data written in the storage unit is based on once every two carry signals output from the read address generator when reading of the image data having a size equal to M blocks is completed. And a step of reading from a memory block obtained by shifting a memory block to be read for each image data by M blocks based on the read address signal.

Images